Generate initial set of auto-inserted links

[DanilaFe]

This PR builds on https://github.com/chapel-lang/chapel-blog/pull/281 to insert documentation links. In this PR, I do the following things:

• Write a new script, check_doc_links.py, which validates the links in the cache. In particular, it loads the pages and tries to find the corresponding fragment of the page.
• Make adjustments to insert_links.py to avoid emitting links that go nowhere (caught by check_doc_links.py). In some cases this is due to custom wrangling that we do.
• Use the updated insert_links.py to generate an initial set of links-to-insert.
• Adjust the CSS to look better.

To address some of @bradcray's concerns, this PR also adds a CI check to verify that links in the link cache are valid.

[bradcray]

@DanilaFe : Q: How hard would it be to make the links to the docs be relative rather than absolute? Wondering from the perspectives of:

• testing when the website's down (doesn't happen often, but is biting me today)
• having the link checking potentially be done completely locally without involving a web server
• permitting the ability to changes to docs and/or blogs before pushing docs changes to the live site

[DanilaFe]

It should be pretty easy. Right now, I have logic that hardcodes the chapel-lang.org URL, but I'm sure I can adjust it for a predefined relative path like ../../. For local link checking, though, not sure if just switching to relative paths would be enough. Probably need to make the link checker script be aware that we are using file:// URLs.

[bradcray]

I'm pretty sure some of the desktop-based link checkers I've used will make sure that an <a href = "../../foo/bar.html"> link is valid, complaining if it's not. I can double-check, though.

[bradcray]

Oh wait, I'm realizing you're probably talking about the link checking you've set up, not a more general link-checking script…?

[DanilaFe]

Yes, I'm referring to the link checking script I'm checking in with this PR (and which would run with CI after this PR merges)